negotiator.boaframework.opponentmodel
Class AgentLGModel

java.lang.Object
  negotiator.boaframework.OpponentModel
      negotiator.boaframework.opponentmodel.AgentLGModel

public class AgentLGModel
extends OpponentModel

Adaptation of the opponent model used by AgentLG in the ANAC2012 to be compatible with the BOA framework. Note that originally, the model sums up all value scores which entails that the total preference of a bid can be as high as offered bids * issues. The value score was equal to the amount of times the value was offered divided by total amount of bids offered by the opponent. In my implementation I normalize each value score by the highest value, similar to the implementation of preference profiles in Genius. Finally I sum all results and divide by the amount of issues. This is identical to assuming that the issue weights are uniform. Tim Baarslag, Koen Hindriks, Mark Hendrikx, Alex Dirkzwager and Catholijn M. Jonker. Decoupling Negotiating Agents to Explore the Space of Negotiation Strategies

Author:

Mark Hendrikx

Field Summary

Fields inherited from class negotiator.boaframework.OpponentModel

negotiationSession, opponentUtilitySpace

Constructor Summary

AgentLGModel()

Method Summary

void	cleanUp() Removes references to the objects used by the opponent model.
double	getBidEvaluation(Bid bid) Determines the utility of a bid according to the preference profile.
java.lang.String	getName()
UtilitySpace	getOpponentUtilitySpace()
double	getWeight(Issue issue) Returns the weight of a particular issue in the domain.
void	init(NegotiationSession negotiationSession, java.util.HashMap<java.lang.String,java.lang.Double> parameters) Initialize the opponent model by creating an object to keep track of the values for each issue.
void	updateModel(Bid opponentBid, double time) Update the opponent model by updating the value score for each issue.

Methods inherited from class negotiator.boaframework.OpponentModel

endSession, getIssueWeights, init, isCleared, loadData, setOpponentUtilitySpace, setOpponentUtilitySpace, storeData, updateModel

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

AgentLGModel

public AgentLGModel()

Method Detail

init

public void init(NegotiationSession negotiationSession,
                 java.util.HashMap<java.lang.String,java.lang.Double> parameters)
          throws java.lang.Exception

Initialize the opponent model by creating an object to keep track of the values for each issue.

Overrides:

init in class OpponentModel

Parameters:

negotiationSession - reference to the state of the negotiation

Throws:

java.lang.Exception

updateModel

public void updateModel(Bid opponentBid,
                        double time)

Update the opponent model by updating the value score for each issue.

Specified by:

updateModel in class OpponentModel

Parameters:

opponentBid -

time - of offering

getBidEvaluation

public double getBidEvaluation(Bid bid)

Description copied from class: OpponentModel

Determines the utility of a bid according to the preference profile.

Overrides:

getBidEvaluation in class OpponentModel

Parameters:

bid - of which the utility is calculated.

Returns:

utility of the opponent's bid

getWeight

public double getWeight(Issue issue)

Description copied from class: OpponentModel

Returns the weight of a particular issue in the domain.

Overrides:

getWeight in class OpponentModel

Parameters:

issue - from which the weight should be returned

Returns:

the uniform issue weight

getOpponentUtilitySpace

public UtilitySpace getOpponentUtilitySpace()

Overrides:

getOpponentUtilitySpace in class OpponentModel

Returns:

utilityspace created by using the opponent model adapter.

cleanUp

public void cleanUp()

Description copied from class: OpponentModel

Removes references to the objects used by the opponent model.

Overrides:

cleanUp in class OpponentModel

getName

public java.lang.String getName()

Overrides:

getName in class OpponentModel

Returns:

name of the opponent model.